Overview of Object Detection Evaluation Metrics

How to Measure the Accuracy of Object Detection Models?

When evaluating the performance of an object detector, we use two main evaluation metrics FPS (frame-per-second) to measure the network detection speed, and mAP (mean Average Precision) to measure the network precision. In this article, we will go through each of them discussing what they mean and how to calculate each of them.

Table of Contents:

1. FPS to Measure the Detection Speed

2. mAP to Measure the Network Precision
   2.1. Intersection Over Union (IoU)
   2.2. Precision-Recall Curve (PR Curve)